细粒度自适应框架的生成机制

Generative mechanism of fine-grained self-adaptive framework

自主计算引起了极大的关注,现在已经有很多此类的项目。例如,密西根州立大学的"TRAP/J"、东京技术学院的"Open Java"、伊利诺斯大学的"PCL"等。参考"TRAP/J",提出了一种细粒度自适应框架的生成机制。只为原来程序中需要自适应能力的部分构筑框架,因此产生的代码简洁、可读性好。并且为用户提供了一次机会,让其决定代理和原来程序之间的关系。这些改进使自适应框架能够更加灵活地适应环境。但是在该项目中,被处理类的一些结构信息需要在程序参数中指定,要求用户十分清楚地了解该类。

Autonomic computing have drawn great attention,and now there are already quite a number of projects,such as ＂TRAP/J＂ of Michigan State University,＂Open Java＂ of Tokyo Institute of Technology,and ＂PCL＂ of University of Illinois,etc.We mainly make a reference to ＂TRAP/J＂,and propose a generative mechanism of fine-grained self-adaptive framework.We just construct the adaptive frameworks for parts of the original program which really need the ability,and so the generated code is brevity and well-readable.And we provide the opportunity for user to decide the relationship among delegations and the original program;it will make the framework more flexible to adapt the situation.However in this project,some structure information of the dealt class need be written in program arguments and it requires the users to comprehend the class clearly.